Re: Is ARMENIAN ABBREVIATION MARK (՟, U+055F) misclassified?
On Sat, 27 Apr 2019 00:08:52 + James Kass via Unicode wrote: > On 2019-04-26 11:08 PM, Doug Ewell via Unicode wrote: > > This is a small percentage of the number of fonts that have all > > four of these Armenian glyphs, but show the abbreviation mark as a > > spacing glyph. It looks like Unicode is right, Wikipedia is right, > > and the fonts are wrong. > If the Wikipedia page(s) are correct, then Unicode isn't. Unicode > charts don't show the glyph on the dotted circle and the canonical > combining class is shown as "spacing". The fact that Doug Ewell > found some installed fonts displaying the character as a combining > mark suggests that the Wikipedia pages are correct. This character > is listed as being unused in modern Armenian, but you'd think that it > would have been exposed before now since the charcter has been in > Unicode since version 1.0. Well, ccc=0 is entirely permissible for non-spacing marks, though I find it an invitation to misspell words. I think the most important admissible issue is one of word boundaries. U+055F has line_break=alphabetic, but word_break=other. The latter doesn't seem very friendly towards spell checkers, but perhaps there is a good reason for it. Word_break=other is not compatible with being a non-spacing mark. Another important, but probably inadmissible, issue is that of the effect on editing. Life is easier if one can easily change the character preceding the abbreviation sign; this would often be difficult if the abbreviation sign were a combining mark. What advantages would accrue from changing U+055F from Po to Mn? Richard.
Re: Is ARMENIAN ABBREVIATION MARK (՟, U+055F) misclassified?
On 2019-04-26 11:08 PM, Doug Ewell via Unicode wrote: This is a small percentage of the number of fonts that have all four of these Armenian glyphs, but show the abbreviation mark as a spacing glyph. It looks like Unicode is right, Wikipedia is right, and the fonts are wrong. If the Wikipedia page(s) are correct, then Unicode isn't. Unicode charts don't show the glyph on the dotted circle and the canonical combining class is shown as "spacing". The fact that Doug Ewell found some installed fonts displaying the character as a combining mark suggests that the Wikipedia pages are correct. This character is listed as being unused in modern Armenian, but you'd think that it would have been exposed before now since the charcter has been in Unicode since version 1.0.
Re: Is ARMENIAN ABBREVIATION MARK (՟, U+055F) misclassified?
Fredrick Brennan wrote: > Although my research on this has by no means been exhaustive, it > seems at a cursory glance that the «pativ», the Armenian abbreviation > mark, is misclassified; it seems it should either be itself a > combining mark or have a combining mark version. > > I have not been able to find a single Unicode font which treats it as > such, however. Using BabelPad on Windows 10, with the sequence <0531, 0532, 0533, 055F> (ayb, ben, gim, abbreviation mark), the following fonts show the abbreviation mark correctly over the gim: Calibri Cambria Cambria Math Nishiki-teki Trebuchet MS All of these except Nishiki-teki are standard Windows fonts. This is a small percentage of the number of fonts that have all four of these Armenian glyphs, but show the abbreviation mark as a spacing glyph. It looks like Unicode is right, Wikipedia is right, and the fonts are wrong. -- Doug Ewell | Thornton, CO, US | ewellic.org
Is ARMENIAN ABBREVIATION MARK (՟, U+055F) misclassified?
Although my research on this has by no means been exhaustive, it seems at a cursory glance that the «pativ», the Armenian abbreviation mark, is misclassified; it seems it should either be itself a combining mark or have a combining mark version. I have not been able to find a single Unicode font which treats it as such, however. Is Wikipedia, then, correct to say «The pativ was used as an Armenian abbreviation mark, and was placed on top of an abbreviated word to indicate that it was abbreviated.»? The image included was built with a LaTeX hack according to its description to get it to build "properly". How should fonts treat this? Should I file a bug in Noto Sans Armenian, and other open source fonts which include U+055F? Or is Unicode right, and Wikipedia wrong? (If so, someone should fix that.)